belforte
commented
10 years ago origin of problem understood: user has a filter in the cmssw config file, note the line: filterName = cms.untracked.string('Digi.root')
process.RAWSIMoutput = cms.OutputModule("PoolOutputModule", splitLevel = cms.untracked.int32(0), eventAutoFlushCompressedSize = cms.untracked.int32(5242880), outputCommands = process.RAWSIMEventContent.outputCommands, fileName = cms.untracked.string('REDIGI_DIGI_L1_DIGI2RAW_HLT_PU.root'), dataset = cms.untracked.PSet( filterName = cms.untracked.string('Digi.root'), dataTier = cms.untracked.string('GEN-SIM-RAW') ) )
this triggers a "never used before" piece of code in ModifyJobReport which inserts the filter name after the numerical has in the processed dataset name https://github.com/dmwm/ProdCommon/blob/master/src/python/ProdCommon/FwkJobRep/ModifyJobReport.py#L261
which fails the Lexicon and anyhow creates an odd situation where pieces of LFN are not matching block/dataset pieces as usual etc.
user managed to insert in DBS3 files like '/store/user/lperrini/05_03_14_higgsToZAllXX_DIGI/higgsToZAtolltautau_350_15/lperrini/higgsToZAToTauTau_mH_350_mA_15_i2HDM_V0_GEN_SIM/higgsToZAtolltautau_350_15_V0_DIGI/2c5f164c0ecfd520d5d2801602656805/REDIGI_DIGI_L1_DIGI2RAW_HLT_PU_1_1_lSv.root'
while the dataset is: '/higgsToZAToTauTau_mH_350_mA_15_i2HDM_V0_GEN_SIM/lperrini-higgsToZAtolltautau_350_15_V0_DIGI-2c5f164c0ecfd520d5d2801602656805-Digi.root/USER'
with the extra -Digi.root
this dataset name can not be use in input to DBS3 API's.
I think the best way to fix is to get rid of that part of ModifyJobReport and stick to ProcessedDatasetName = user-PublishName-psethash
since in any case we will not support multiple output datasets in Crab2
as a side note, I have asked DBS3 to prevent the publication of those kind of blocks, so damage is limited next time something like this happens: https://github.com/dmwm/DBS/issues/368
see Stefano Belforte
11:09 AM (1 hour ago)
to hn-cms-crabFee. *\ Discussion title: CRAB Feedback
thanks Lucia. Indeed also the log where you managed to publish shows the same error: 2014-03-06 17:17:56,593 [INFO] Error when listing files in DBSHTTP Error 400: Invalid Input Data /higgsToZA...: Not Match Required FormatTraceback (most recent call last):
And I see a very odd dataset name: /higgsToZAToTauTau_mH_350_mA_30_i2HDM_V0_GEN_SIM/lperrini-higgsToZAtolltautau_350_30_V0_DIGI-2c5f164c0ecfd520d5d2801602656805-Digi.root/USER
the "-Digi.root" before /USER should not be there.
I can't say yet if this is a wrong configuration on your side or a bug in the crab pre-release.
Please send me the crab.cfg you are using for the failing crab -create.
stefano
On 03/07/2014 10:55 AM, Lucia Perrini wrote:
to hn-cms-crabFee. *\ Discussion title: CRAB Feedback
thanks Lucia. Indeed also the log where you managed to publish shows the same error: 2014-03-06 17:17:56,593 [INFO] Error when listing files in DBSHTTP Error 400: Invalid Input Data /higgsToZA...: Not Match Required FormatTraceback (most recent call last):
And I see a very odd dataset name: /higgsToZAToTauTau_mH_350_mA_30_i2HDM_V0_GEN_SIM/lperrini-higgsToZAtolltautau_350_30_V0_DIGI-2c5f164c0ecfd520d5d2801602656805-Digi.root/USER
the "-Digi.root" before /USER should not be there.
I can't say yet if this is a wrong configuration on your side or a bug in the crab pre-release.
Please send me the crab.cfg you are using for the failing crab -create.
stefano
On 03/07/2014 10:55 AM, Lucia Perrini wrote: